The Agent Skills Directory

[COMMAND_EXECUTION]: Uses the gh CLI (including gh repo view, gh issue list, gh label list, and gh issue create) and grep to interact with GitHub repositories and local files for backlog management tasks.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it ingests and processes untrusted data from external sources.
Ingestion points: Data is read from docs/BACKLOG.md, individual markdown files in docs/issues/, and the output of the gh issue list command.
Boundary markers: There are no explicit delimiters or instructions to ignore potential commands embedded within the retrieved issue content.
Capability inventory: The skill can write to the local file system (docs/BACKLOG.md, docs/issues/) and create/modify issues on GitHub via the gh CLI.
Sanitization: The instructions do not specify any validation, escaping, or filtering of the content read from files or CLI output.
Mitigation: A strong mitigation is present in the instructions, which explicitly require 'per-item confirmation' from the user before performing destructive or state-changing actions like closing or merging items.

backlog